Data Analysis

Fracture orientation was estimated by applying different methods to the different data sets available, started by analyzing the P-S converted waves in the multicomponent data. Then, we analyze the AVO and NMO responses of P- waves recorded in the vertical component of the 2-D 3-C data around the intersection point of the three lines.

Finally, the azimuthal variations of AVO and NMO responses of the P-waves for each bin in the 3-D data was investigated. The following sections describes the obtained results.

2.4.1. 2-D data

Rotation analysis of P-S converted waves.—Figure 19 shows portions of migrated horizontal components around three different points in the field. Notice that points located on lines 1 and 3 clearly show that the horizontal component parallel to line 3 arrives earlier than the other component, which is what we expected from the direction of the maximum horizontal stress in the field.

Remember that the orientation of the faster shear arrival generally coincides with the direction of maximum horizontal stress.

Figure 14. Maximum horizontal stress (inward facing arrows) from break-out orientation logs at wells, 16, 17, 20, and 23. The rose diagrams indicate fracture orientation and density from FMS logs in the same wells (Perez et al, 1999) .

Figure 17. 3-C raw common-shot gather recorded over line 2 of the 2-D multicomponent lines. Converted waves from the top of Escandalosa Formation are indicated by the arrow at 3.8 s in the horizontal components (Perez et al, 1999).

Figure 18. Structural map of the top of Escandalosa Formation interpreted from the 3-D seismic data. Colors indicate two-way traveltime in second (Perez et al, 1999)

After interpreting the migrated sections of radial and transverse components, rms amplitudes in a window around the target was compute. Then, rotational analysis based on the amplitude ratio between the two horizontal components (Ata and Michelena, 1995) to estimate fracture orientation for each common conversion point (CCP) of the three lines was performed. From the angles estimated at each CCP, we obtain new angles for points outside the multicomponent lines using 2-D spline interpolation. The orientation of the fastest shear arrival happens to be approximately constant for all depths across the field.

Figure 20 shows a smoothed map with the results of the rotational analysis plus interpolation.

Each arrow indicates the local fracture orientation (fracture strike) estimated from converted waves. The colors show the same structural map presented in Figure 18. Since the azimuths between the lines have been interpolated, orientations presented in this map are more reliable along and in the neighborhood of the three lines. As we can see, fracture orientation follows the trend of the maximum horizontal stress in the area. Notice that the estimated fracture orientation follows the direction of one of the fracture systems present in the area.

Azimuthal AVO from 2-D data.—For the 2-D P-wave data, we perform conventional AVO analysis over CDP gathers located along each line close to the intersection point.

We obtain the sections of AVO gradient and AVO intercept for each line. Figure 21 shows the AVO gradient sections around the intersection point along almost perpendicular lines 1 and 3. Notice the different AVO gradients for the reflection from the bottom of the reservoir along these two lines. No significant differences are observed in AVO response from the top of the reservoir. Figure 22 is a graph of AVO gradient versus AVO intercept. Gradients for lines 1 and 2 (nearly perpendicular to fracture orientation) are positive and higher than gradients along line 3 (which is nearly parallel to the fracture orientation estimated from converted waves). As expected, the AVO intercept is almost the same for all lines.

The exact direction of the maximum AVO gradient was estimated from the three CDP gathers located at the intersection point of the multicomponent lines. We use a formula derived by Rüger (1996) for the reflection coefficients in transversely isotropic media with a horizontal symmetry axis. The estimated azimuth of the maximum AVO gradient is 56°. Since it is expected to be perpendicular to fracture orientation, we obtain the fracture azimuth is 146° (Figure 23) .

Fracture orientation estimated using this technique also follows the regional maximum horizontal stress and is close to the orientation obtained from the analysis of the converted waves (Figure 20).

NMO ellipticity from 2-D data.—From the same three gathers used to obtain the orientation of the maximum AVO gradient, we estimate the parameters that describe the best fitting horizontal ellipse of the NMO velocities for all azimuths (Grechka and Tsvankin, 1996). The NMO ellipse obtained from this analysis is shown in Figure 21. The azimuth of the major axis is 125°. The major axis corresponds to the maximum NMO velocity that is expected to coincide with fracture orientation.

The differences between maximum and minimum velocities are around 5%, but the axes of the ellipse have been exaggerated to give a better idea of its orientation. In this case, the orientation of the major axis of the NMO ellipse also follows the regional trend of the maximum horizontal stress in the area.

2.4.2. 3-D Seismic Data

Azimuthal AVO from 3-D data.—3-D P-wave data was

gathered using a bin size of 240 x 240 m2 _{to achieve the coverage}

needed in both offset and azimuth to perform azimuthal AVO analysis. The orientation of the maximum AVO gradient was estimated for each superbin based on the amplitudes located within a time window that follows the top of Escandalosa.

Figure 25 shows smoothed results of the azimuthal AVO analysis for each superbin. The arrows in Figure 25 are oriented according to these new, smoothed angles. As we can see, the estimated orientations follow closely the local structural changes, but the general trend is still close to the regional maximum horizontal stress. Areas with abrupt changes in structure seem to have a more erratic AVO response when compared to other areas in the field.

Notice the similarity between the results obtained with converted waves (Figure 20) and 3-D azimuthal AVO analysis (Figure 23). There are differences, however, between the two results, especially in the northwest part of the area.

No interpolation was used to generate Figure 12 because AVO analysis was performed at equally spaced grid points. On the contrary, to generate Figure 7, we interpolated the azimuths measured along the lines for all grid points where no information was available.

In principle, this can create unrealistic orientations in areas surrounded by angles that represent the same direction but different orientations (0º and 180º, for instance ), which is the case in the northwest part of the area. No attempt was made to change these angles in the data before interpolation.

Figure 21. AVO gradient sections around the interception point for

lines 1 and 3. Notice how the AVO responses changes around the bottom of the reservoir located at 2.370 s (Perez et

Figure 22. AVO intercept versus AVO gradient for lines 1,2 and 3 at the intersection point of the 2-D multicomponent lines (Perez et al, 1999)

Figure 23. Orientation perpendicular to the maximum AVO gradient at the intersection point of the 2-D multicomponent lines. This result is compared to the orientation of line 3, which is nearly parallel to the orientation of maximum horizontal stress in

Figure 24. Orientation of the NMO ellipse from 2-D P-wave data at the intersection point of the 2- D multicomponent lines (Perez et al, 1999).

Figure 26. Fracture orientation from 3-D NMO ellipticity. The arrows indicate the local orientation of the maximum axis for the NMO ellipse (Perez et al, 1999).

NMO ellipticity from 3-D data.—3-D NMO analysis was performed for the same superbins used in the 3-D azimuthal AVO analysis. Figure 26 shows the result of this process: the orientation of the semimajor axes of NMO ellipses at the target for each CDP. The general trend in the orientation of these axes is not as expected from all our previous results, which followed the orientation of the maximum horizontal stress more closely. In Figure 26, the differences between the semimajor and semiminor axes of the NMO ellipse are of the order of 3%.

We can speculate about various reasons to explain the orientation of NMO ellipses for each CDP. First, the NMO ellipses include a cumulative influence if the overburden from the surface to the target.

We did not do layer stripping to obtain interval velocities because the target is too thin (25 m) compared to its depth and, therefore, the error amplification due to stripping is expected to be severe. Second, since the ellipticities (i.e., the elongations of NMO ellipses) are small (about 3%). The ellipse orientation becomes a poorly determined quantity, so that the estimated ellipse azimuths may be inaccurate. The third issue is that we believe that the proper way to remove the effects of near-surface azimuthal anisotrophy is by doing independent static corrections for each azimuth, which we did only for the 2-D data. The effect of static variations, near- surface anisotropy,

and azimuthal anisotropy in the subsurface that affect the NMO velocities cannot be separated whit static correction methods that analyze simultaneously all azimuths in a superbin. We found that after applying such corrections for any depth, lateral coherency of events was improved considerably, but differences between major and minor axes of NMO ellipses were reduced to less than 0.01 %.

For this reason, we did not remove static corrections in the data used to generate Figure 27. Finally, the influence of lateral heterogeneity may make azimuthal variation of the NMO velocity elliptical even in the absence of anisotropy.

Figure 27 shows lateral variations of the isotropic NMO velocity ( e.g., a circular approximation of NMO ellipse) estimated from conventional velocity analysis around the target. Even though the changes in these velocities are not that large, they may distort our inferences about P-wave azimuthal anisotropy which are made under the assumption that the medium is laterally homogeneous.

BIBLIOGRAPHY

Alford, R. M., 1986, Shear data in the presence of azimuthal anisotropy: 56th Ann. Internat. Mtg., Sac. Expl. Geophys., Expanded Abstract, 476-479.

Ata, E., and Michelena, R. J., 1995, Mapping distribution of fractures in a reservoir with P-S converted waves: The Leading Edge, 12,664-676.

Corrigan, D., Withers, R., Damall, J., and Skopinski, T., 1996, Analysis of amplitude versus offset to detect gas-oil contacts in the Arabian Gulf: 66th Ann. Internat. Mtg., Soc. Expl. Geophys., Expanded Abstract, 1834-1837.

Garotta, R, and Granger, P. Y., 1988, Acquisition and processing of 3C x 3-D data using converted waves: 58th Ann. internat. Mtg., Soc. Expi. Geophys., Expanded Abstract, 657-658.

Grechka, V., and Tsvankin, J., 1996, 3D description of normal moveout in anisotropic media: 66th Ann. Internat. Mtg., Soc. Expl. Geophys., Expanded Abstracts, 1487-1490.

Johnson, W. E., 1995, Direct detection of gas in pre-Tertiary sediments?: The Leading Edge, 14, 119-122. Lefeuvre, F., 1994, Fracture related anisotropy detection and analysis: and if the P-waves were enough?": 64th Ann. Internat. Mtg., Soc. Expl. Geophys., Expanded Abstracts, 942-944.

Li, Y. et al, 2003, Recent application of AVO to carbonate reservoirs in the Western Canadian Sedimentary Basin, The Leading Edge, July 2003, vol.22, no.7, SEG.

Lynn, H., Simon, K. M., Layman, M., Schneider, R, Bates, C. R., and Jones, M., 1995, Use of anisotropy in P- wave and S-wave data for fracture characterization in a naturally fractured gas reservoir: The wading Edge, 14, 887-893.

Mallick, S., and Frazer, L N., 1991, Reflection/transmission coefficients and azimuthal anisotropy in marine seismic studies: Geophys. J. Internat., 105,241-252.

Maria A. Pérez*, Vladmir Grechka‡_{, and Reinaldo J. Michelena*, 1999, Fracture detection in a carbonate}

reservoir using a variety of seismic methods, Geophysics, v.64, no.4, 1266-1276

Michelena, R. J., Ata, E., and Sierra, J., 1994, Exploiting P-S converted waves: Part I, Modeling the effects of anisotropy and heterogeneities: 64 Ann. Internat. Mtg., Soc. Expl. Geophys., Expanded Abstract, 236--239. Perez, M., and Gibson, R., 1996, Detection of fracture orientationing azimuthal variation of P-wave AVO responses: Barinas Field (Venezuela): 66th Ann. Internat Mtg., Soc. Expl. Geophys., Expanded Abstracts, 1353-1356.

Perez, M., Gibson, R., and Toksoz, N., 1999, Detection of Fracture orientation from azimuthal variation of P- wave AVO responses: Geophysics, 64, 1253-1267, this issue.

Ruger, A., 1996, Reflection coefficients and azimuthal AVO analysis in anisotropic media: Ph.D. thesis, Colorado School of Mines.